review

Author: mcb5637

private/rdf4cpp/regex/RegexImpl.cpp

@@ -114,10 +114,10 @@ namespace rdf4cpp::regex {
         auto s = make_code(regex, flags, 0);
         return {std::move(m), std::move(s), flags};
     }
-    Regex::Impl::Impl(code_ptr m, code_ptr s, flag_type f)
-        : match(std::move(m)),
-          search(std::move(s)),
-          flags(f) {
+    Regex::Impl::Impl(code_ptr match, code_ptr search, flag_type flags)
+        : match(std::move(match)),
+          search(std::move(search)),
+          flags(flags) {
     }
 
 }  //namespace rdf4cpp::regex

private/rdf4cpp/regex/RegexImpl.hpp

@@ -8,6 +8,7 @@
 
 namespace rdf4cpp::regex {
     struct Regex::Impl {
+    private:
         // workaround for gcc-14 bug, erroneously warns on unsing a lambda here
         // see https://github.com/NVIDIA/stdexec/issues/1143
         template<auto f>
@@ -18,6 +19,8 @@ namespace rdf4cpp::regex {
         };
 
         using code_ptr = std::unique_ptr<pcre2_code_8, CallFree<pcre2_code_free_8>>;
+
+    public:
         code_ptr match;
         code_ptr search;
         Regex::flag_type flags;
@@ -34,7 +37,7 @@ namespace rdf4cpp::regex {
         static code_ptr make_code(std::string_view regex, flag_type flags, int extra_flags);
         static std::string remove_whitespace(std::string_view str);
         static Impl make(std::string_view regex, flag_type flags);
-        Impl(code_ptr m, code_ptr s, flag_type f);
+        Impl(code_ptr match, code_ptr search, flag_type flags);
     };
 }  //namespace rdf4cpp::regex
 

private/rdf4cpp/regex/RegexReplacerImpl.cpp

@@ -4,19 +4,19 @@
 #include <uni_algo/conv.h>
 
 namespace rdf4cpp::regex {
-    std::string RegexReplacer::Impl::translate_rewrite(std::string_view const s) {
+    static std::string translate_rewrite(std::string_view const s) {
         std::string res{s};
 
         auto pos = res.find_first_of("$\\");
         while (pos < res.size()) {
-            if (res[pos] == '\\') {  // NOLINT(*-pro-bounds-avoid-unchecked-container-access)
-                if (res.size() <= pos) {
-                    throw RegexError{"invalid escape sequence in replacement string"};
+            if (res[pos] == '\\') {
+                if (res.size() <= pos + 1) {
+                    throw RegexError{"incomplete escape sequence in replacement string"};
                 }
-                if (res.at(pos + 1) == '$') {
-                    res[pos] = '$';  // NOLINT(*-pro-bounds-avoid-unchecked-container-access)
+                if (res[pos + 1] == '$') {
+                    res[pos] = '$';
                     ++pos;
-                } else if (res.at(pos + 1) == '\\') {
+                } else if (res[pos + 1] == '\\') {
                     res.erase(pos, 1);
                 } else {
                     throw RegexError{"incomplete escape sequence in replacement string"};

private/rdf4cpp/regex/RegexReplacerImpl.hpp

@@ -12,9 +12,6 @@ namespace rdf4cpp::regex {
 
         Impl(std::shared_ptr<Regex::Impl const> regex, std::string_view rewrite);
         void regex_replace(std::string &str) const;
-
-    private:
-        static std::string translate_rewrite(std::string_view s);
     };
 
 }  //namespace rdf4cpp::regex

src/rdf4cpp/regex/RegexFlags.hpp

@@ -12,6 +12,7 @@ enum struct RegexFlag : uint8_t {
     Literal = 1 << 2,
     Multiline = 1 << 3,
     RemoveWhitespace = 1 << 4,
+    // enables extra optimization during the parse process (currently JIT-compiling), leading to an increase in regex compile time, but faster matching & replacing
     Optimize = 1 << 5,
 };
 

tests/util/tests_Regex.cpp

@@ -197,6 +197,8 @@ TEST_SUITE("regex") {
         CHECK(d == "b_b_x");
         Regex{"_(\\w)", f}.make_replacer(R"(-$0$1\$\\)").regex_replace(d);
         CHECK(d == "b-_bb$\\-_xx$\\");
+
+        CHECK_THROWS_WITH_AS(static_cast<void>(Regex{"a", f}.make_replacer("\\")), "incomplete escape sequence in replacement string", RegexError);
     }
     TEST_CASE("basic syntax") {
         SUBCASE("normal") {